Ladies and gentlemen, welcome to the ATOSS Software SE Q4 and full year 2025 earnings call. I am Matilde, the call operator. I would like to remind you that all participants will be in listen-only mode, and the conference is being recorded. The presentation will be followed by Q&A session. You can register for questions at any time by pressing Star and One on your telephone. For operator assistance, please press Star and Zero. The conference must not be recorded for publication or broadcast. At this time, it's my pleasure to hand over to Carla Leicher, Manager, Corporate Development. Please go ahead.
Thank you, operator, and hello, everyone. Welcome back to our third earnings call, where we will be discussing our results for Q4 and full year 2025. We are pleased to have you here with us today. I am joined by our Chief Financial Officer, Christof Leiber, and we are glad to have the opportunity to walk you through our performance and outlook. We will be referring to the earnings call Q4 and full year 2025 presentation, which was published earlier this morning and is available for download on our investor relations website, as well as via the link provided in the webcast. A detailed investor relations presentation was also published this morning, which we encourage you to review for further insights, but will not discuss during this call.
Please note that today's call is being recorded, and the recording will be made available on our investor relations website after the call. Before we begin, I would like to start with a disclaimer. Please note that the presentation contains forward-looking statements based on the beliefs of ATOSS Software SE. These statements reflect the current views of ATOSS Software SE with respect to future events and results and are subject to risks and uncertainties. Actual results may differ materially from those projected here due to factors including, but not limited to, changes in general economic and business conditions, the introduction of competing products, lack of market acceptance of new products, services or technologies, and changes in business strategy. ATOSS Software SE does not intend or assume any obligation to update these forward-looking statements.
With that, I will now hand over to Christof Leiber, who will walk you through the key developments of the fourth quarter and full year of 2025. He will start with a general business update, then cover our financial performance, followed by our outlook for 2026 and beyond. We will then wrap up with a Q&A session. Christof, over to you.
Thank you, Carla, and a very warm welcome from my side to all of our to the entire audience. I'm happy to be here today and walk you through our Q4 figures as well as our full year 2025 results. We appreciate your time and interest in ATOSS, and look forward to sharing an update on our performance and outlook, including our AI roadmap beyond 2026. Let's get started on slide 4 with key takeaways. We are proud to announce that 2025 marks our twentieth consecutive record year. 20 years of continued year-on-year growth, top line, and in terms of EBIT. Our CAGR for the years since 2014 has been at 15%. Overall, this is a unique success that makes everyone at ATOSS extremely proud.
We also successfully completed our midterm guidance that we gave on the back of 2022 for the years 2023-2025. For this period, we projected a CAGR of approximately 19%. Now, we came in at 18.5% CAGR for that period. Also, we projected an EBIT margin by the end of 2025, above 30%. Now, our EBIT margin is two years in a row, above 35%. You've seen the numbers presumably already this morning. So ATOSS once again has demonstrated its exceptional success as a software company and our delivery ability on demanding guidances. This is success that I like to dedicate to our employees, our customers, and our clear commitment to our vision. At the same time, we've built a platform at ATOSS that is extremely well-positioned to continue on this path in the future.
But let us first focus on the results for 2025. We've delivered solid results with full year revenues of EUR 189.3 million, and an EBIT margin of 36%, well above guidance and reflecting the quality of our earnings. As for the growth within our medium-term guidance, once again, the growth has been driven by cloud and subscription revenue streams. This, this sort of revenue was up 28% for the full year of 2025, and we closed the year really on a high note. Q4 came in strong with 12% top line growth, quarter-over-quarter, driven by, once again, strong momentum in cloud and subscriptions, which was equally up 28% in Q4 as for the entire year. I mentioned last quarter that comparables in Q4 had been expected and actually were tougher, and still the team has delivered.
Many thanks to everyone at ATOSS for making this possible. On the order intake side, we've seen a substantial improvement in H2 2025 over H1 2025, and we believe that Q1 and H1 2026 should show a continuation of this positive development. In particular, we are pleased with the development on the new ACV generated for cloud and subscription. Here we saw strong growth comparing to H2 2020, comparing H2 2025 to H2 2024, and still good growth comparing full year 2025 to 2024. Again, cloud and subscription remain our main growth engine, supported by strong order backlog, healthy ARR momentum, and a solid new ACV development. These are the key areas to determine, from our perspective, the value of ATOSS. Now, building on this momentum, we enter 2026 with confidence.
We confirm our outlook of around EUR 215 million in revenue and an EBIT margin of greater than 32%. As always, our revenue target is realistic, yet ambitious, with a bandwidth of roughly 2%. Our EBIT margin and target is conservatively set, meaning we have leg room to either outperform, cover the lower end of the revenue bandwidth or or/and initiate additional investments. For 2026, we expect cloud and subscription revenues to grow by 25%+, and total recurring revenues by more than 15%, while on-prem licenses will continue to represent a smaller and more volatile share of the mix. Also, for 2027, we stand by our projection of around EUR 245 million, with a bandwidth of 3% and an EBIT margin of at least 33%.
Now, on slide 5, let's take a look back at the development of the phenomenal 20-year record history that we at ATOSS have achieved. Since 2014, ATOSS has even increased its growth momentum with an average annual rate of 15%. This was substantially driven by strong development in recurring revenues, in particular since 2021. Today, our recurring revenues account for 94% of software revenues and 70% of total revenues, and this share continues to rise. Internationally, we keep, kept expanding, averaging a 30% growth over the past 5 years. In 2025, revenue outside of the DACH region accounted for 6% of total revenues. More importantly, we have won great new customers internationally in 2025, and in particular in H2 2025.
Two large retailers in Benelux, one retailer in the Middle East, and somewhat as the icing on the cake, one international retailer headquartered in France in Q4. Moving on to margins. Margins have been substantially increasing during 20, during our 20 record year trajectory, and practically, in all years, we outperformed our guidances on margins. As you know, we tend to keep our margin guidance conservative, and we will keep it this way. Now, with that, let's move to the income statement on slide 6 for the full year 2025. We have delivered a solid year 2025, keeping the track we were on in Q4, Q3 2025, with revenues increasing by 11% year-on-year, driven by software revenue growth of 13%.
As said, cloud and subscription business continues to be the major driver of this growth, with 28% year-over-year increase in cloud and subscription revenues, which now account for nearly 50% of our total revenues, up from 42% last year. For other revenue streams, consulting showed a continued, good growth at 10% year-over-year, and other revenues, including process consulting, even stronger growth of 31% year-over-year. Whereas hardware revenues and perpetual licenses, as already alluded to, saw a decline in 2025. On the margin side, we stayed strong despite the investments in our go-to-market organization and its now completed transformation. Looking at Q4 alone, and on slide 7, top line growth was again strong at 12% year-over-year.
Software revenues grew at a similar rate of around 12%, keeping both metrics broadly in line with the levels of the previous year. Overall, this was a solid quarter in terms of order development, with particular strength in on the cloud and subscription side in order intake, revenues, and overall, very strong margins once again. This provides a good support for our revenue development going into 2026. Taking a closer look at our cloud and subscription and overall recurring revenue, ARR performance on Slide 8, we continue to see a very positive development. Our total ARR grew by 18% to EUR 140 million last year. This trend is largely driven by very strong momentum in our recurring revenues streams, cloud and subscription.
ARR increased by 28% year-on-year to EUR 101 million by the end of Q4 2025. Looking further into the drivers of our cloud and subscription ARR, it is important to highlight that the majority of the ARR growth in 2025, again, came from new business rather than migrations. Out of the EUR 22 million ARR increase year-on-year, more than 80% came from new ARR with existing and new customers. With existing and new customers evenly split, and only 20% or roughly nearly four million euros came from migrations. This demonstrates the externally driven momentum of our cloud offering and confirms the attractiveness of our solution in the market.
When looking at customer value dynamics, our net retention rate came in at around 111% for 2025, and as I mentioned already multiple times and some time ago, our long-term ambition has always been an NRR above 110%. That is where we are, and that has been expected from the outset as our cloud customer base over time becomes increasingly larger, and the share of newer customers with more appetite for expansions of the cohort decreases. Now, coming to our backlog, which gives us good indication of future ARR development, our total ARR backlog increased by 18% year-on-year, reflecting the strong level of contractually committed additions for the next 12 months. In addition, the incremental cloud and subscription backlog added in 2025 increased. You see that on this slide.
It has increased after two years of rather flattish environment. So this shows a positive change in the last year, with 8% growth of that incremental cloud and subscription backlog added in 2025. Both indicators further strengthen our visibility into 2026 and underline our continued robustness of our recurring revenue model. Together, these developments provide a solid foundation and support a positive outlook for our ARR development in 2026. Now, turning to cash flow and liquidity on slide 9. Operating cash flow in 2025 came in lower year-on-year. This decline does not reflect weaker underlying operating performance by no means. It is essentially driven by exceptionally high tax cash outs in 2025.
Because of additional taxes set after the final tax assessment for 2023 and additional prepayments for 2024, all in total tax payments in 2025 amounted to EUR 28.4 million versus EUR 9.8 million, so roughly EUR 10 million in 2024. When normalized for the additional tax payments for 2023 and additional prepayments for 2024, the operating cash flow would have increased year-on-year in 2025. Now on the liquidity side, we remain very, with a very strong position. Even after dividend payment of around EUR 34 million in Q2, we closed the year with a liquidity level of around EUR 123 million, slightly above EUR 112 million at year-end 2024. This strong foundation allows us to continue investing in the business while maintaining a very healthy financial posture. Moving on to people and organization on slide 10.
At the year end 2025, our total headcounts stood at 856, compared to 820 at year end 2024. This reflects a moderate, largely planned expansion of our organization, which is overall still in line with our transformation efforts. With regard to our go-to-market organization, we are clearly on track. At the year end, our sales and marketing headcount stood at 201 employees, which is within the range that I was giving on the last previous calls, between 200 and 210. The particular focus area continues to be our quota-carrying organization. At the end of Q4 2025, when combining the staff on board and the already hired staff with entry date up to April 1, approximately 70 quota carriers and first-line managers will be on board.
While this is slightly below the target of around 80, we are making steady progress and remain confident that we will close this gap very soon. In parallel, we continue to drive efficiency improvements across our go-to-market setup through better processes, AI, and digital tools, ensuring the productivity rises overall. Overall, this gives us a significant, stronger commercial engine entering into 2026, with more structure, more capabilities, and greater resilience against macro fluctuations. With a strong finish in Q4, on slide 11, we close out 2025 as our 20th consecutive record year. For full year, we delivered revenues of more than EUR 189 million, fully within our guided range, and achieved an EBIT margin of 36%, clearly above our already raised guidance of 34% EBIT margin. On the recurring side, we saw a continued strong momentum.
Total recurring revenues grew by around 18% year-on-year, and cloud and subscription revenues increased by 28%, both in line with our expectations. As mentioned earlier, the risk in the model continues to be limited, mainly on to the on-prem licenses and a bit on hardware, which represent a very small fraction and a very limited share of our total business. Looking ahead to 2026, we confirm our guidance. We expect revenues to come in around EUR 250 million, meaning a 2% bandwidth, and in continuation of our conservative margin guidance practice, an EBIT margin of greater than 32%. This outlook is supported by strong ARR and backlog in cloud subscription. And EUR 245 million by 2027, representing a CAGR of roughly 14% from 2025 onwards. Naturally, the pace of growth will depend on macroeconomic conditions and our sales execution.
But overall, the progress we made in 2025, we feel well positioned for the years ahead. Looking toward 2030, we believe that leveraging our position and the market dynamics for workforce management, we can build an organization of around or with around EUR 400 million in revenue by 2030. In this direction, we continue to invest to even increase our current organic growth, and we are opening up to selective inorganic growth. But of course, it is an ambition and not a guidance. Now, moving to slide 12. One area, among others, for our future growth is continued investment in our AI roadmap. ATOSS has started the AI journey in 2022, 2023 with industry specialists like University Medical Center Mainz for hospitals, Fraunhofer Institute, and others, which this led to the initial set of AI services.
In 2024, 2025, ATOSS started delivery of AI services with a clear focus on forecasting, as this is the base for any optimization of workforce scheduling. This has already positively impacted the development shown in our ARR order backlog, in particular for cloud and subscription in 2025. For hospitals, there are substantial advantages in faster assuring the right forecast and therefore, thereby building planning decisions on this accurate and faster and readily available forecast. This is applicable, of course, not only for hospitals, but also for the rest of the industries as well. The topics here are general forecasting, illness rate forecasting, as well as vacation or other absence forecasting. That is what we have already delivered last year.
However, the AI roadmap is obviously much more, and as you see on this slide here, it is key to the entire roadmap of ATOSS for all of our solutions. Our focus on AI agents creating workforce efficiency, productivity, and simplification is clear. We cater to end users, employees, and experts, and thereby covering the full breadth of the workforce. This year, in 2026, we will have the first AI agents for the expert users and employees across the product lines. Furthermore, we have a multi-year roadmap with focus industries touching upon every aspect of workforce management. On our roadmap are multiple agents for ASES, configuration agents, voice agents for multiple use cases, for managers, casual users, et cetera. Agents for specific industries like retail, just to name a few off this list here.
Multiple agents equally for ATC, AI assistants to speed up and automate the planning process, AI support for forecasting, et cetera. Business intelligence also for ATC. And even for Crewmeister, we have multiple agents on the roadmap, AI-based crew administration, as well as AI-based scheduling suggestions, et cetera. Now, there's a lot to come, and we are just getting started. From our perspective, success in AI is based on three pillars, if you will. Domain expertise and workforce management is an area with a huge moat for ATOSS that is rooted in this domain expertise. Deep embeddedness in customer processes and understanding, for example, of the regulatory compliance and complexity. How else should an AI leverage the potential thereof? And lastly, access and ownership of the data that is relevant to workforce management.
Now, ATOSS, we are bringing all of this and combine it with our AI innovations. This will further deepen our existing moat in workforce management, and still has much more potential. Now, I guess this concludes the presentation part of the call. We'd now like to open the floor for questions, and are happy to dive deeper into any topics that you would like to discuss. Thank you.
We will now begin the question and answer session. Anyone who wishes to ask a question may press star and one on their telephone. You will hear a tone to confirm that you have entered the queue. If you wish to remove yourself from the question queue, you may press star and two. Questioners on the phone are requested to disable the loudspeaker mode, and eventually turn off the volume from the webcast while asking a question. Anyone who has a question may press star and one at this time. The first question comes from the line of Nicolas Herms from Deutsche Bank. Please go ahead.
Yeah. Hi, Christof. Thank you for having me, and congratulations on another record year. I have a couple of questions, but I would like to start with sort of the obvious question on the risk from AI. The market appears to be pricing in on application software these days. So yeah, I mean, you've gave some color on that already, but it would just be interesting to hear your perspective on, you know, potential risks you are seeing on your business and how these developments impact your customers. And most importantly, do you see any of your customers starting to, you know, try different tools to respond to needs that were previously addressed by ATOSS?
And then, on the other hand, maybe on the opportunities from AI and the product update that you have given, just a couple of minutes before, can you maybe share some initial customer feedback from the features that you have already launched, or that you have, you know, announced before? And, could you maybe remind us how you're planning to monetize these features? Thank you.
Well, thank you, Nicolas, for the question that you just raised, and obviously, AI is on everybody's agenda right now. Let me start with first an observation. I think everyone got really excited last, and I think wrongfully so in some respect, as last year in August, or so it was, I just came back from vacation, Sam Altman put out that claim that the death of software is around. Now, first of all, ATOSS will not be beaten or eaten by this. And secondly, I think it's rather the other way around. It is, AI obviously is extremely transformative. It will be helping customers to become more efficient, but efficiency is based on three pillars, if you will, and I try to kind of make these points out.
One is that you need to have domain expertise. And in our core area, that is domain expertise of workforce management. If the vendor of AI solutions or services has no clue of what workforce management is about, and it's an extremely complex scenario or topic, then it is extremely difficult. You can code very fast with AI, but you need to code meaningful. Now, so this is the domain expertise. The second thing is, our solutions and workforce management solutions of ATOSS with based on this big moat that we have built over the years, are deeply embedded with our customers.
Now, imagine a company like Deutsche Bahn, for example, or Lufthansa, Deutsche Bahn, with more than 100,000 employees live, and they actually last year decided to expand very substantially into ATOSS going forward. Now, imagine these processes for multitudes of different companies in such an organization. They can leverage AI, but they can only leverage it if you can really deliver end-to-end digitization, and that is what we are very much about, at least in our enterprise area. And this then leads me to the third pillar, which is the data access.
In order to really fully make leverage AI, and we've learned that in our forecasting AI service, you need to have access to meaningful data, and and make use of this data in order to come up with credible, reliable forecast prognoses, but just faster and more adaptable, as in the old days, where you did have to parameterize a lot of stuff. Now, our AI forecasting service is doing amendments, et cetera, more or less automatically. So that's why I'm not really seeing the threat. I'm rather seeing the potential that there is for companies like ATOSS with a clear moat, with a deep embeddedness within customers, and then bringing AI services on top of our solutions, that is actually creating value.
Now, you were asking about how this value is being seen, and how we can probably speak a bit about making it visual, that people understand that already AI is delivering a positive impact on our numbers. Now, I cannot give you a clear number, but what I can give you is the indication that last year, all of our hospitals, the university hospitals and other hospitals that we've won, and we've won quite a few of them, they have selected ATOSS for a multitude of reasons, obviously. But one of the reasons was that we have an AI agenda, that we have AI services for forecasting already readily available, and that made them choose ATOSS. That made them choose ATOSS as a cloud solution, because then you'll get access to the AI services.
And this, in particular, in a market environment where you do have—and when speaking about workforce management, the market environment is a bit different than in other areas. You have very small vendors, really. You don't have large vendors, for the most part. You have smaller vendors. This creates and entrenches, really, the moat that ATOSS has developed over the past years. And last year, lastly, you asked the question about whether we encounter already customers making use of independent, let's say, large language model providers, allowing or delivering AI services to these, customers. That is not, not what we encountered so far. Maybe in some, areas where there's very little complexity, but I have not heard about this at this point.
As I said, we believe that rather we can leverage the benefits of AI. Obviously, we need to stay innovative, and then this will be a positive for us, clearly.
Thank you. That's, that's very helpful, and actually what we hear in our customer discussions as well. I have a quick follow-up, if I may, on the revenue guidance of around EUR 215 million. I, I recall you previously mentioned that achieving the EUR 215 million would require roughly flat order intake in 2025, which you have now delivered. So I was wondering, what are the drivers or assumptions that would lead you to the upper or the lower end of that guidance range?
Yeah. Very valid question. Now, first, our order development last year, I would want to make this point, I really had to split the year in two halves. The first half, 2025, was not a strong half. We had externally difficulties. I mean, the macroeconomic situation was not entirely good. It didn't really improve throughout the year. In Germany, at least, we had negative sentiments by terrorists, at least impacting our potential customers. So the first half was externally not good, and quite frankly, internally, we had a lot of things to do. We had the transformation of our go-to-market organization still very much ongoing. We had to change our CRO during this time. So first half was not very good. The second half was extremely good, by comparison.
In the second half, if you compare H1 with H2, H2 was significantly outperforming H1, more than double-digit growth there. And if you look for the full year on the cloud and subscription development there, as I said, we had a strong performance year-on-year for cloud and subscription, so 25 versus 24. And here, again, if you compared H2 2024 with H2 2025, there will be a strong uptake. This is showing slightly in the incremental order backlog for cloud growth, which is now a solid growth, I would say, or good growth, as I refer to it, between 5%-10%, and that's roughly what we increased in cloud and cloud order subscription. So we are actually on a good trajectory there.
This leads us to the ARR backlog of 146.5, which if we add to this EUR 43 million, roughly, of consulting revenues, plus EUR 10 million of other and hardware, and then EUR 9 million, roughly, of perpetual licenses, we end up with 209, and so without any new hire, new cloud contracts being signed. So with new cloud contracts, we just barely need EUR 5-6 million in new ACV or new revenue next year, which should be possible. Here, as I said, the cloud, on the cloud side, on the recurring revenue side, we do have very good visibility. The limitations in our visibility still come from the perpetual side, where we do see basically two ways this could go.
One way would be the longer trend, a continuation of the longer trend, which ultimately I would see, which would mean a further slight decline. That would pose a risk, a slight risk on our guidance. That's why we came out with the bandwidth of 2%, to the lower end. The upper end, obviously, could be that with all this talk about sovereignty in, in Europe or in, in Germany, that we do have the, yeah, the ability at least, to see some more perpetual in a short period of time, while this sovereign talk will push some customers to the cloud. So broadly speaking, we are seeing the EUR 215 as a realistic, yet ambitious guidance, 2% bandwidth, and the risk is with the perpetual licenses.
All right. Thank you. Thank you very much. Very helpful.
The next question comes from the line of Philipp Sennewald from NuWays AG. Please go ahead.
Hi, Christof. Thank you for the presentation, and congrats also from my side. You mentioned the order momentum has caught up significantly in the second half of the year. I would be interested in your perspective, is this only a catch-up effect, or is this genuinely stronger underlying demand in your view?
Yep. Philipp, thanks for the question, and I try to make the point that we see this as from an organizational point of view, from the dynamics of the market, in particular in the public sector in our area, as something which has the chance to continue, and actually not just the chance, we've in particular, in H1 and Q1, we would envision that our order development will stay on this path. For the second half of the year in this kind of environment, I cannot really project clearly, but at least for Q1 and H1, based on the development that we're seeing, some deals that could have been closed in Q4 but moved to Q1.
So we have quite a good pipeline for Q1, and H1 should be on top of last year as well. That then would hopefully tie into the further maturity of our sales organization with newly hired people adding to positive effects in the second half of this year. So that overall, for this year, we are quite optimistic in starting on a quite optimistic turn.
Thank you. That's, that's very helpful. Next one would be on Crewmeister. Crewmeister showed a slightly weaker net retention this year than last year. What were the reasons here? How do you aim to stabilize it, and do you have a long-term target for Crewmeister regarding net retention?
On first of all, on Crewmeister, I think Crewmeister has added significant customer numbers this year as well. We had hoped initially for a slightly higher number. We ended up nearly at 18,000, so I think it's 17,900 or so that we came in with. We had hoped for above 18,000, so slight decline. Again, it was also due to, from my perspective, H1, which came in lower. We reassessed the ways in which we reach out to customers in this area, opened up new channels, relaunched our website, and enhanced traffic there as well. So the second half was quite stronger, so we actually had there as well a second half, which was pleasing.
Going forward for this year, we envision the customers to increase to 22,000. Precisely on net retention, well, I mean, this is a different business altogether, so it's always a bit... Yeah, I think maybe not okay if we add this up in our overall net retention, which still with Crewmeister stands at 111%. We are working on churn. We have improved a bit, the churn, but it still stands at 1.5, I think, roughly per month, the churn there, which is our struggle, or which is our kind of key point, which we need to improve in order to move up the gross retention, and thereby then the net retention, as well.
Still a very dynamic area, a lot of potential, and yeah, we have to work a bit on the churn side.
Yeah, perfect. Thank you very much. That also helps me a lot. And then one last, you have EUR 13.1 million in your cloud subscription ARR from new and migrated customers. Can you distinguish there, what of that is new and what of that is migrated customers?
Yes. I tried to do this in my presentation, but once again, there I gave the number that 80% of the 22 in total comes from new licenses, from existing and new customers. They are evenly split, but to give you the precise numbers of the 13.1, 3.9 are coming from migrations. The remainder, 9.1 or one-- 9.2 or so, 9.2, comes from new logos. So that's quite pleasing, seeing that 50% basically of the ARR expansion in the cloud and subscription side is coming from new logos. And 50% basically comes from the existing customer side, and then added 20% from internal customers, if you will. So maintenance customers migrating into the cloud.
Oh, perfect. Thanks. I'm sorry, I didn't get that, in the presentation. That's it from my side. Thank you again.
Thank you.
We now have a question from the line of Gustav Froberg from Berenberg. Please go ahead.
Good afternoon. Thank you for taking my questions as well. Just a couple. First on uptake and success of the new product. I mean, you mentioned it a little bit, but could you tell us a little bit more about the uptake you've seen with some of the new products and features you've launched, in 2025? And maybe give us an indication as to which industries are particularly active on taking up new solutions. Then a question on the agentic AI product you're rolling out for Q2 of this year. Are there any other similar products in the market today, or do you think that you are very early or first to market with something like this?
And then lastly, just on, on migrations, et cetera, how should we think about the migration momentum into 2026? Do you expect migrations to accelerate, or are you making any, concerted effort to push, for more migrations, or are you expecting the pace to be rather, as it has been in the past? Thank you.
Okay. Thanks, Gustav. And thanks, by the way, for the very, very deep review that you put out, I think just a short while ago on workforce management. Maybe it's something worth looking at for others as well. Now, taking your question, uptake of the current services. Now, the uptake of the current services, which are forecasting services, has mainly been taken, and then the feedback that we got was in hospitals. As I said in my presentation, we initiated the launch, the innovation of our forecasting AI services with the University Hospital of Mainz.
We built the prototype, so it's all geared towards this medical environment, and there it has significantly impacted the deals that we have won since, I would say, end of 2024. And basically, all of these deals had, to a certain fraction, this element of, we want the AI service for forecasting. So in that sense, it has already delivered quite a positive effect on our order development and on our ARR development, but I cannot quantify this at this point. And we will, however, and it's perfectly possible to use it in other industries as well. This will be something which we probably have to educate people a bit more about and kind of go out more into the market.
But generally, forecasting and AI-driven forecasting is always the better way of doing forecasting than the old way that we had done with the classical AI, as I called it for some time. But it's the classical way was algorithm-based and not machine learning like the new AI service. So, and the classical way is basically installed in all of our retail customers, and they would tremendously benefit from moving to the AI service going forward. Now, to your question on the more imminent new service that we will bring out in Q2 for ATC, the agentic use cases there, in particular for the SMB product, ATC. I'm always stressing this point that overall, the market for workforce management is a very fragmented market.
That holds true from enterprise to the very low end to the micro company markets, where Crewmeister is active. But obviously, in SMB, the lower you get, the more fragmented it is. So in this area, you have competitors that are extremely small, very small, and by bringing out an agentic AI use case, it definitely sets ATC apart from the other vendors that are out there and the competitive products. So in that sense, we hope that this definitely will bring an uptake, a continued uptake. We had a good year in SMB in 2025 as well, order intake-wise, but we hope to kind of see even more thereof in 2026, among other things, based on the agentic use case that we have there for ATC in Q2. Lastly, on the migration side, there we are following a twofold strategy.
I think that has not changed. I think I've discussed it in the Q3 earnings call as well, which is, simply put, for the enterprise customers, we stay true to our commitment to our customers of continuously delivering on-prem, in particular, in an environment where on-prem customers, enterprise customers, are looking for sovereign solutions. Among others, there are sovereign cloud solutions as well, but there's one angle of sovereignty that they can and will continuously get from us with our on-prem offering there. So this is the starting point. Secondly, we are inviting, and we will give incentivize, and with the new AI services, we do have tools to incentivize these customers, the on-prem customers, to migrate to the cloud faster than they probably would have otherwise.
So we will leverage these new toolbox, if you will, that we have, moving our customers that are currently on-prem and moving them into the cloud. We will package this together with economically interesting offerings for them to limit the uplift that they would have to pay for the migration, and then make continuous value, and we, of course, would hope for the share in that value that our customers are then generating. So that's the other side. No force, but incentives in, in moving, and thereby, I would see a slight uptick in the migration there, but not a substantial one because we don't force. On the ATC side, it's slightly different.
There, we do have the clear focus and clear view, or at least ambition, to migrate all of our customers by 2030, and we will start with a combination of incentivization, creating interest for the AI services that we now start to deliver in 2026, firstly, to ATC customers. But we will add to this some sort of economic push as well, if you will. We will start to enhance or increase maintenance costs by sometime this year or at beginning of next year, and announce it this year, that maintenance costs for ATC will rise. So there's more economic sense in moving to the higher value cloud solution with embedded AI use cases, so there will be value on this side as well.
At some point, there will be an end there as well for the ATC side. So that's kind of the two-way approach. This two-way approach will mean that our customers that we hope for a slight uptick, but I wouldn't count on... We don't plan for substantial changes in the dynamic that we've seen in 2026 or 2025, sorry.
Great. Thank you very much.
Thank you.
As a reminder, if you wish to register for a question, please press star and one on your telephone. The next question comes from the line of Gustav Froberg from Berenberg. Please go ahead.
Hi again. Sorry, just a follow-up question. Just on new features again and monetization of those, are you thinking about monetizing them in the form of standalone, pricing, perhaps, on a per token basis or per use basis? Or are you looking to bundle them, as part of the existing solution and some kind of upsell, motion? I'd just be curious on the pricing strategy. Thanks.
Yeah. Thank you for that question, and very valid one, obviously, because there is different dynamics with AI services as they require some computing power as well and create costs, obviously, as well. Now, with these forecasting services, currently, we embed them or sell them as a separate module. So this is more the add-on or the effect of winning more customers and enhancing the overall ticket volume for the particular customer. Now, going forward for the agentic use cases, we will envision a hybrid pricing structure where we do have then, on one hand, token-based computing power-related price methodology, and of course, in parallel, the similar subscription-based pricing as we do have today.
It will be a combination of both. Currently, for the current modules, the forecasting ones, we have not implemented that, but for the new agentic AI services, it will be implemented this way.
Great. And is there any gross margin difference between what it is that you envisage to charge for customers on a token basis versus a subscription, or not really?
It's too early to tell, really, but I would, I mean, our goal is to keep our healthy gross margins stable, and I don't wanna go into the details of where our gross margins stand at this point.
Okay. No problem. Thank you very much.
Ladies and gentlemen, that was the last question. I would now like to turn the conference back over to Christof Leiber for any closing remarks.
Well, thank you, and thank you for all the attendees and your interest, your continued interest in ATOSS. Just in a nutshell, I just want to reiterate, the entire team here at ATOSS is extremely proud to have delivered 20 consecutive years. And not just because we have... Basically, we're always looking in the back mirror, and then we, we celebrate ourselves for having the 20 years. The 20 years success is really about having built a platform that gives us comfort to look into the future. In this area of workforce management, we have the financial means, the innovation capabilities, and a lot more that will put us in a position to leverage the opportunities that there are with AI for a software vendor that has deep embeddedness with our customers, good exchange and continued exchange with our customers.
And that is what we hope we will leverage, not just in 2026, but all the way to 2030, and there are so many more things that we could talk about, but I hope we leave that to the next earnings calls. And with that, I'll close it, and thank you for your interest. Thank you.
Ladies and gentlemen, the conference is now over. Thank you for choosing Chorus Call, and thank you for participating in the conference. You may now disconnect your lines. Goodbye.